While Artificial Intelligence is the center piece topic in the realm of information technology, its less sexy counter part – boring ho-hum “Automation” – is having just as big an impact on our economy, on jobs and our civilization overall.
Currently, we encounter “automation” everywhere: at the grocery story in the self-service check out lane, at the airport with the check-in kiosks, and now even at Mc Donalds with ordering kiosks. The routine tasks of: (1) adding up every item purchased by a store cashier is now done by the customer in the self-service check out lane of a grocery store, (2) checking a passenger into a given airline flight is now done by the passenger at a kiosk and (3) ordering a McDonalds Happy Meal is now also done by the customer at a kiosk. These routine tasks have been automated as customer self-service tasks. These companies are taking these tasks away from employees and they are assigning these tasks to their own freakin’ customers as “self-service” tasks! What Chutzpah!!! There is usually no AI involved with automating these tasks (at least for now!). These routine tasks have been transformed through plain vanilla automation.
Consider these two blog posts. The first is by Andrew Yang in Chapter Six of his book War on Normal People. The following excerpt is freely available from Yang’s web-site. The link is below immediately after the following excerpt :
“We tend to think of automation as displacing blue-collar workers with jobs that involve basic, repetitive skills. The truth is a little bit more complicated than that. The important categories are not white-collar vs. blue-collar or even cognitive skills vs. manual skills. The real distinction is routine vs. non-routine. Routine jobs of all stripes are those most under threat from AI and automation, but in time more categories of job will be affected. Doctors, lawyers, accountants, wealth advisors, traders, journalists and even artists and psychologists who perform routine activities are threatened by automation technologies in the coming years. Some of the most highly-educated jobs are actually among the most likely to become obsolete. Some of these threatened workers, like investment advisors, may find themselves surprised to be on the chopping block after supporting the profit-growing potential of automated technologies.”
The Jobs That Will Be Lost Are Not What You Think
Notice how Yang uses “both” the terms (1) artificial intelligence and (2) automation. It is very important to know the similarities and differences between these two terms.
This Yang blog post is followed up with a post on using AI as a supporting tool in non-routine jobs. In the following blog post, the author also uses both the terms “artificial intelligence” and “automation”:
“I also specify nonroutine tasks – the work that isn’t highly repeatable or structured and requires the worker to invent how to get the task done as they go. With outsourcing and automation taking more and more of the routine work, nonroutine work is where developed economies generate the most value and differentiation. Because it is tacit, applying simple automation can be difficult, but techniques like deep learning can be applied.” https://blogs.gartner.com/craig-roth/2017/12/05/489/
This subject of “automation” has been the primary focus of my career since 2015. It was in 2015 that I decided to focus 100% on automating as much of the work I had performed for 25 years. Since 1994, I have been working with a range of Cisco networking products. Throughout this same period, I had been automating much of my work with a patch work of tools such as:
Expect
Perl
BASH
SNMP (specifically command-line SNMP)
Virtually all of these tools allowed me to automate the human interaction I had with Cisco devices through an SSH terminal session.
However, around 2008 Cisco began releasing platforms that possessed interfaces that allowed for pure automation and not simply automating the human interaction with the device. These new platforms such as the Cisco NEXUS switch platform and the Cisco UCS offered NETCONF and REST interfaces. Other network equipment vendors such as Juniper Networks and Artista were offering similar 100% automation interfaces.
Also, there was the emergence of Amazon Web Services (AWS). One of the primary benefits of AWS is the high degree of automation is offers. Along with the emergence of AWS is also the development of Google Cloud Platform, the amazing Microsoft corporate-wide transformation of becoming a cloud centered company with Windows-365 and Azure, VMWare’s virtualization offerings and finally, Cisco broad offering of private and hybrid cloud solutions. At the center of all of these corporate offerings from these big tech names is: automation. Terms such as “software defined networking” and “infrastructure as code” underlie and reflect the degree of automation of all of these big name offerings.
All of this has guided the development of: IaaS, PaaS and SaaS.
https://www.bmc.com/blogs/saas-vs-paas-vs-iaas-whats-the-difference-and-how-to-choose/
All of these new services have API interfaces.
My profession was in transition and I wanted to be ahead of the transition and not playing catch up. By 2015, I made the transition to perform as much of my network engineering and support operations via automation. It is now 2019 and my automation tool kit includes:
Python
Open Source IDE’s: Jupyter Notebooks, PyCharm
JSON and XML and supporting tools: jmespath, xpath, etc.
REST and REST tools such as Postman
NETCONF/RESTCONF
Data Modeling Languages (YANG & OpenConfig)
Puppet/Chef/Ansible/SaltStack
Splunk/ELK Stack
git and github
I called this transition to automation: From CLI to API.
While all of the tools and technologies above offer the benefit of eliminating routine tasks through automation, these very same tools and technologies create a new challenge and ultimately a new problem: complexity and required support for this collection of tools. As with virtually all technology, “when you are 99% right with technology, you are very likely 100% wrong”. As your reliance on building out and expanding automation solutions with the technologies above, your time troubleshooting, incrementally enhancing and supporting these tools also expands as well. What can ultimately happen is: you are spending so much time troubleshooting and updating your automation solutions that it might be better to abandon many of these automation solutions and go back to performing some of these tasks manually. Going back to performing tasks manually and abandoning automation???? No, we can let this happen! There are ways to make your automation effort “scale”.
Deploying “scalable” automation, you must embrace the following:
- code sharing using a platform such as git hub
- be flexible and agile be ready to migrate to more scalable solution when they become mature
You just can not sit still and wait for either AI or even simple automation to develop and mature. Taking this approach may result in what one of my students in the Philippines said to me, “Learn to automate or be terminated”. Yes, it’s this serious. You might not be terminated from your job by not keeping up with this automation trend. But you might be made a part-time employee, passed up for a promotion or limited in career development. At the very least, we all must be “literate” with these trends in automation.
We will close with a little story: A few years back AWS experienced and outage due to a glitch in its scalable automation tools. How did AWS overcome this glitch? Did it abandon its automation effort? Of course not, it solved its glitch with…. MORE AUTOMATION!
Let’s all get literate on this topic of automation. Let’s learn about automation’s current capabilities, its current limitations and its trajectory into the future. Automation ain’t goin’ away! Let’s do this together! Let’s Unite with Code!!! We are… UnitedWithCode! : )
